Quality Control and Activity Monitoring in Telecommunication Industry

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

Telecommunication industry is a highly competitive industry. Since divestiture and government de-regulations, various telephone services, such as cellular, local and long distance, domestic and commercial, have been the battle fields for telecommunication service providers. To remain competitive, it is important for the companies to establish business planning systems to support management for strategic planning and decision making. Statistical process control (SPC) techniques have been used routinely for on-line process control and monitoring and have resulted in cost-effective returns on investment in manufacturing industry. Our research builds on the powerful SPC and modeling methods to develop efficient statistical techniques for data mining and activity monitoring in telecommunication industry. We identify the limitations of the current SPC methods in telecommunication applications. We develop efficient methods for monitoring large volume of univariate and multivariate data and construct economic models to evaluate and compare various monitoring methods. We also develop design of experiments and churn modeling methods to identify important characteristics responsible for abnormal activities and to predict likelihood that the activities would happen. The major impact of the proposed research is to improve the existing techniques and tools in the business planning system for data mining and knowledge extraction in the telecommunication industry. The enhanced system will allow marketing and sales managers to apply data mining methods routinely, efficiently, and effectively to solve their problems. As a result, our research will help the company reduce cost through efficient fraud detection systems and improve revenue through customer churn prevention, customer retention programs, and new business opportunities based on extracted useful knowledge.